Algorithms for the Time Window Assignment Vehicle Routing Problem
Vehicle Routing Problem, Time Window Assignment, Pharmaceutical industry,Coopservice Company, Hybrid algorithm.
We study the Time Window Assignment Vehicle Routing Problem (TWAVRP), which appears in real contexts where demand is unknown or suffers fluctuations per delivery. We consider a finite number of scenarios, capacitated vehicles, and exogeneous time-windows for each client. Our goal is to minimize transportation costs and to assign endogeneous time-windows overall scenarios. We propose a hybrid algorithm that generates a set of routes by requesting an Iterated Local Search (ILS) metaheuristic and then chooses the most appropriate routes through a set-covering based auxiliary formulation. The contributions described here is threefold. First, we improve the best-known solutions reported in the TWAVRP literature. Then we extend the TWAVRP adding heterogeneous vehicles and driver stopping periods assumptions. This variant appears in the pharmaceutical industry. With a database provided by Coopservice company, we generate a new real instance. Finally, we test our approach with such an instance. Computational results indicate that the proposed algorithm is accurate in practice, and answer our research question: “What are the algorithms that can optimize costs and respect all constraints of TWAVRP and its variant concerning the Coopservice routing planning?'”